Debate on GSM8k
84.4AccuracyOPTIMA-iDPO SC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| OPTIMA-iDPO SCoptimization_strategy=iDPO, self_consistency=true2024.10 | 84.4 | 1,750.8 | |
| OPTIMA-iSFT-DPO SCoptimization_strategy=iSFT-DPO, self_consistency=true2024.10 | 84 | 1,748.7 | |
| OPTIMA-iSFT SCoptimization_strategy=iSFT, self_consistency=true2024.10 | 83.1 | 1,750.7 | |
| OPTIMA-iSFT-DPOoptimization_strategy=iSFT-DPO2024.10 | 80.4 | 246.5 | |
| SCn=82024.10 | 80.3 | 1,828.7 | |
| OPTIMA-iSFToptimization_strategy=iSFT2024.10 | 79.5 | 311.5 | |
| OPTIMA-iDPOoptimization_strategy=iDPO2024.10 | 78.5 | 270.1 | |
| MAD2024.10 | 72.5 | 514.7 | |
| CoT2024.10 | 71.5 | 230.9 | |
| AutoForm2024.10 | 71 | 410.5 |